Using NetFlow to Improve Network Visibility and Application Performance

Editor's Notes

• #9: You may not be familiar with SevOne, so before we get into the ways in which Emulex and SevOne use NetFlow to improve visibility of your infrastructure’s health, perhaps a quick introduction…
• #11: Our goal is to help you avoid business disruption due to issues with your infrastructure performanceWe do this in two ways…
• #12: First, we help you Prove it– SevOne proves your ability to deliver business critical applications and services on a consistent basis.Are you meeting challenging SLAs? When there’s an issue, is it a network issue, a server issue, a storage issue?Can you prove the success of a new technology deployment, like that datacenter virtualization project or VoIP deployment? To what degree did it improve the performance of your infrastructure?We also help you Predict it – service impacting events have a significant impact on your organization’s revenue, and the longer it takes to repair the issue, the more costly the impact, so SevOne proactively alerts you to potential issues with your infrastructure before they impact your customers or end users. If you’re in capacity management, our customers also use SevOne to determine future planning needs.[pause] so that’s SevOne, in a nutshell. And we believe we are the fastest, most scalable performance management platform on the market today… primarly because our customers have told us so.
• #13: The single common thread with all of our customers when we first speak with them is that they have some degree of visibility gap when it comes to the performance of their infrastructure.Maybe you’re a wireless operator and the deployment of new small cells in your LTE environment has resulted in lost visibility of network performance…Or maybe that new virtualization project has created visibility gaps in regards to how the guest and host machines interact with each otherI recently saw a study by TRAC’s research that revealed that – for organizations who recently conducted virtualization projects -- visibility into the performance of IT services declined 38%, on average61% of organizations reported that their visibility into the quality of user experience declined after deploying public cloud services. So lack of visibility = more headaches for you, because it means you’re unable to provide a direct and complete answer when your manager or exec or anyone else asks…Why did we experience poor quality on our VoIP call this morning?Why was our web site taking a long time to load pages over lunch?Why was our load balancing effort not evenly distributed across the servers in the pool?
• #14: SevOne provides those answers by incorporating traditional performance metrics -- like SNMP and IP SLA -- with NetFlow recordsSNMP will tell you that something occurred, like spike in utilization, but to answer the “who” and “what” questions, you would turn to NetFlow for deeper insightThis is not a separate module or tool… it’s part of the all-in-one SevOne solutionIf there’s a spike in bandwidth utilization, from the same dashboard you can see who the top talkers were at the time the spike occurredWe’ll talk more about how the joint deployment of SevOne and the Emulex Endace solution increases performance visibility, but first…
• #15: Just so you better understand the value that SevOne drives for you, I want to talk briefly about why our customers say they like usBecause I think the reasons our customers like us ties directly back to the product requirements and recommendations that Lee touched on at the end of his presentation
• #16: You can’t watch everything all of the time in your environment… you need a solution to automate that process for you… to alert you when performance metrics start to stray from what would be considered expected behavior.
• #18: If you ask any organization, at least half of them are going to admit that they are still struggling with: 1) troubleshooting performance issues in a timely manner; 2) proactively preventing performance problems.
• #19: [iPhone 5 deployment -  identified the key devices that executed SIM card activations – built a dashboard to monitor the activation host performance and session counts, synthetic response time tests, firewalls and load balancers… everything in the data path… to see if anything ever began to deviate from normal]One of our customers, Cablevision, recently reported back to us that 30% of all the alert they receive from SevOne are proactive notifications of issues that are correctable before significant impact occurs.
• #20: [iPhone 5 deployment -  identified the key devices that executed SIM card activations – built a dashboard to monitor the activation host performance and session counts, synthetic response time tests, firewalls and load balancers… everything in the data path… to see if anything ever began to deviate from normal]One of our customers, Cablevision, recently reported back to us that 30% of all the alert they receive from SevOne are proactive notifications of issues that are correctable before significant impact occurs.
• #21: Capacity trends proactively alert teams as to the next resource to reach max utilization and future projections reveal how much time capacity planners have until upgrades will be required. How can you consolidate under-utilized resources and accurately determine future needs of the business?No one likes to have to throw more money at the problem, so how can you make better use of what you have?
• #22: Ease of installation is another consideration that Lee spoke to…
• #23: SevOne is an all-in-one applianceThemonitoring, data collection, alerting, and reporting are all done by a single machineSupport for all of the monitoring and reporting technologies you need are included,out of the box and ready to go
• #24: No security updates or configuration utilities to worry about
• #25: In fact, Comcast, one of our other customers, said they saw an 80% decrease in operational expense after migrating to SevOne (which leads us to the next reason our customers like SevOne)
• #26: Why are you able to achieve such a significant reduction in OpEx, like Comcast did?
• #27: Primarily because SevOne’s underlying architecture allows the solution to scale much ore cost effectively than other solutions on the market. A single SevOne appliance… [explain]
• #28: It would take a “leading” vendor enough servers to stretch the length of a Boeing 737 to monitor 4.2 million elementsThat’s a lot of hardware
• #29: On the other hand, SevOne can monitor more than 4 million elements from a single server rackThis is a killer solution for service providers or anyone looking to limit their hardware footprint and power consumptionYour data center real estate is valuable
• #30: And finally, Lee talked about support for 3rd party applications and 3rd party data sourcesSevOne isn’t always the source of the data, but because of our ability to scale with the largest of networks and still generate report in seconds, we are often the platform of choice for consolidating all performance metrics onto a single pane of glass, as they say.Basically, you can bring any time-based third party data into SevOne for baselining, alerting, and reporting and analysisA common example might be temperature data… [explain]
• #31: In the case of Emulex, our partner on the call today, they’re able to do some things with NetFlow and data collection that goes beyond SevOne’s capabilitiesFor example, while SevOne can perform high frequency polling of objects down to the second – which is far superior to what our completion can do – we don’t go sub-second. This is where the Emulex solution comes into place, by being able to detect and alert on sub-second performance and microbursts that you might not otherwise have visibility of, and then present unsampled NetFlow records in the SevOne dashboard or report engine, which you can see here.
• #32: At the end of the day, it all comes back to, “How can I achieve better visibility of my infrastructure at a lower cost?”